HTINews Feature Article
|Which One Should I Use - Part II
Controlling Motors and Transformers
by Phil Kingery
To all of you who enjoyed and benefited
from my other articles, thank you for your nice emails. This one is a continuation of my
first "Which One Should I Use?" If you have not read the first one, it is a good
prerequisite to this installment: Use the index at the left to go to the December Issue or
use this link (http://www.hometoys.com/htinews/dec96/articles/kingery/kingery.htm).
"Some non-linear loads just should not be controlled with a dimmer at all. Even when the dimmer is full on, it is still not a true sine wave and therefore causes the load to hum, buzz, run hot, or worst case, burn out."
That one sentence motivated many of you to
email asking for more information. Well, get comfortable, here it is.
There is an interesting and often confusing aspect of home automation, and that is, "speed" control. I am constantly amazed by just how easily we all associate "dimming" with "speed control". On more than one occasion , when I have had a particularly bad day, I have been asked about using dimmers for ceiling fan speed control, to which I sarcastically respond, "Yes, when you dim a ceiling fan, it begins to disappear until eventually you can't see it at all. You can feel the breeze but you don't know where its coming from." (Most people don't seem to appreciate my sarcasm.) Let's begin by looking at the basic differences between dimming and speed control.
As you probably know, incandescent lights are resistive loads, which means that, electrically speaking, they have no (or insignificant amounts of) inductive or capacitive reactance. In other words, as far as the electricity is concerned, incandescent lights "feel" like big resistors. More importantly, as standard X-10 dimmers are concerned, incandescent lights "feel" like big resistors and not capacitors nor inductors. Most modern dimmers, including all X-10-based dimmer units, are triac based. A triac dimmer does not attenuate the 60 Hz sine wave, as we might expect.
If we were to use a big variable resistor or a variable transformer, the entire sine wave could be reduced in amplitude (see figure 1). Variable resistor dimming is what was used in stage lighting back in the first half of this century. This, however, is a very inefficient and expensive method of dimming. Today, we have solid state dimmer units that will reduce the light's output not by attenuating the entire sine wave but instead, by quickly turning the sine wave "on and off". Think of it like this. If you were fast enough to flip a switch on and off again, so that only a part of each half-sine wave was allowed to pass through to the light bulb, it would glow only a fraction of its normal brightness. That is what the triac does. Instead of being a variable resistor, it cuts out varying sized chunks of the sine wave. To really see what the output of a triac dimmer looks like you need to look at it like an oscilloscope.
Figure 2 shows a couple of sine waves as they would appear on the output side of a standard X-10 dimmer. You will notice that the leading edge of every half-sine wave has a chunk removed from it. Interestingly, even when the X-10 dimmer is "full on", the light is never really at 100%. Instead it is only allowing about 96% of the power to flow through it to the light bulb. You may have noticed (or thought it was just your imagination) that as soon as you installed your X-10 dimmer, the light doesn't seem as bright as it used to be. There is a very good reason for that. It isn't. It is about 4% lower than it would be with a standard mechanical switch. This is done by design.
It is common, especially in older homes, to find that there are only two wires in the wall box. That means the circuit was originally wired with the line and neutral wires going from the breaker panel to the load junction box, and then two-conductor wire (with ground wire, I hope) extending from that box to the wall box where the switch was installed. Those two wires are not line and neutral, they are really line and switched-line (or hot and switched-hot, colloquially speaking). In order to have a dimmer that will work in this instance, most X-10 dimmers (and those from Radio Shack) have only two wires and are meant to operate on the trickle current "through" the light bulb's filament to power the dimmer circuitry (see figure 3).
Since incandescent lights are "linear" loads, the two-wire X-10
dimmer receives a steady and "linear" amount of current on which to operate.
When the light appears to be "off", the filament still allows enough current to
flow to keep the X-10 unit alive. Even when the light is supposed to be full
"on" the X-10 unit keeps about 4% of the power for itself. If it were to allow
all 100% of the power to flow to the light, there would be none left for its own
circuitry. That is why in figure 2, there is a small chunk of each half cycle left out.
That is the 4% that is being used by the dimmer unit. This also explains why X-10 dimmers
must have a minimum load. Using a 400 or 500 watt load allows for more than enough current
to keep the X-10 circuitry working. Using a 40 watt or 60 watt light bulb will still allow
sufficient current to flow. However, if the load is too small, like a small wattage night
light, then there will be too little current to keep the dimmer unit alive. It will stave
to death (and you don't want that on your conscience, do you).
There is another greater benefit to that 1.1 milli-second gap. That creates a clean space in which the X-10 pulses will appear (see figure 4). Having such an abrupt and instantaneous "switching" action going on twice each cycle has a very definite downside. In that instant in time where the electron flow is trying to go from nothing to full on, that abrupt transition causes a lot of electrical "noise". The X-10 engineers, therefore, designed the X-10 dimmers to always leave the first little section clear so that the units would be able to receive the X-10 signals.
The 3-wire dimmers (those which require their own connection to neutral) really don't need to keep that 4% for their own use, since they have their own source of power (see figure 5). Even so, they will also always leave the first 1.1 milli-seconds blank so that they can receive clear signals.
What really gets confusing is when an X-10
dimmer unit is used on inductive loads. Inductive loads are typically motors and
transformers. Since the output of a triac based, X-10 dimmer is non-sinusoidal (that means
its not really a sine wave), the motor or transformer tends to be "unhappy" with
it. Inductive loads are designed to use clean, smooth sine waves. They don't like it when
their source of power is all chopped up. If you do use an X-10 dimmer with an inductive
load, you will most likely notice that it (the motor or transformer) begins to hum or
Now, lets talk about motors. A standard AC motor relies on the flux lines created around the stator at 60Hz to cause a mechanical rotation in the rotor. A "synchronous" motor is the least tolerant of speed variations due to load. It will attempt to maintain exact speed (in relation to the supply frequency) by drawing more current to increase its torque. Since most AC motors rotate in relation to their supply power, the most widely used method of commercial speed control is the "variable frequency drive". These units are usually abbreviated "VFD's" or "VSD's" (for variable speed drives). These electronic speed control drive units will control the speed of the motor by actually changing the frequency of the power to the motor. To have the motor run slower, the drive supplies power below the standard frequency of 60 Hz. For the motor to run faster than normal, the drive will supply power at a frequency greater than 60 Hz.
Other AC motors will "slip" as the load increases because their available torque can no longer keep up with the demand of the load. Some AC motor controls take advantage of this relationship by limiting the current to a motor, thereby limiting the torque. As the torque is reduced, the motor slows because it simply can no longer maintain its speed. There are also AC/DC motors whose speed can easily be controlled with a simple variable transformer. That is somewhat how an X-10 dimmer unit controls the speed of some motors.
Aside from the receiver section, the X-10 unit is a basic triac type dimmer. Its triac chops out greater and greater chunks of the sine wave as it dims the light. When set for 50% brightness, its triac "holds back" the first half of every half cycle (see figure 6).
And of course, when the light bulb is nearly out, the triac is cutting out nearly all of each half sine wave (see figure 7).
This same reduction in power can be used to control the speed of some motors.
The frequency remains the same although it is no longer a true, smooth sine wave. Ceiling
fans are the most common motorized load for which a home owner (or home automation
installer) wants speed control. Unfortunately, ceiling fans use a wide variety of motors.
Some manufacturers will even use a different type of motor every other month on the same
model fan. Some fans can be controlled by a standard dimmer. It is well known that
"shaded pole" and "permanent split capacitor" type motors are the best
candidates for use with triac based (X-10) dimmers.
There are still some drawbacks to using triac dimmer units (regular or X-10) even with the most compatible of motors. The motor may run hotter than normal since it is no longer receiving a smooth sinusoidal wave as its power source. Being a "non-linear" load, what may be listed as a 1/5hp fan, may appear to be far greater than that to the dimmer triac. For a few milli-seconds each sine wave, the triac may "think" it is connected to a dead short. For those few milli-seconds the triac is being asked to deliver far more current than it was designed to handle, and so it burns up. Even if the triac based dimmer is capable of delivering the current for those short durations, the motor and the associated circuitry inside the fan may not be able to handle it and so they burn up. Don't be discouraged, however. Many times they work fine but we will get into that later.
Now, let's discuss transformers. When someone says "halogen lights" they often mean "low-voltage" halogen lights. In order to tightly wind the filament into a compact space, many decorator lamps used 24v bulbs. That means that the power to run them must come from something that will reduce the regular house power of 120v down to 24v. That used to mean a "transformer". (Now, it may more likely be a low-voltage electronic power supply, but let's talk about transformers first.)
A transformer is, of course, an inductive load and its basic design has not changed much since Edison's day. In simple terms a transformer is a primary coil of wire that allows the flow of current, whose flux lines induce a voltage into a secondary coil of wire. In this case a 120v 60 Hz primary power, induces a secondary 24v 60 Hz power (see figure 8). These coils of wire are wound very close together and are kept apart by "laminates". As long at the transformer is powered by smooth sine waves (remember those back in figure 1), they induce a smooth output in the secondary side. Unfortunately, should the input power "not" be smooth, the transformer may not like it. The same problems that occur when a chopped up sine wave is supplied to a motor also occur when a chopped up sine wave is given to a transformer. Cheap transformers will literally come apart. I have seen a few instances where a triac based dimmer was being used with an old, cheap transformer. The adhesive holding the layers of wire and laminate were so poorly bound that the transformer buzzed like a 2 pound bumblebee. It finally began to tear itself apart. More expensive transformers will do better because their physical construction tolerates the stresses caused by the irregular power.
Even with all of that, there is another problem with many X-10 designed dimmer units (at least in this application). Back in "PART I", we discussed the occasional problem of using a 2-wire dimmer whose supply of power "and" signal must come through the load before it reaches the receiver (see figure 9). Sometimes, a user will notice that a newly installed unit will operate locally but he can not remotely send an "OFF" command to it. Once it is turned off manually, he can send an "ON" command, but once on, he can not get it to go off again. This is caused by the slight change in signal strength seen by the dimmer in the two states. When the light bulb is off, the receiver has nearly the full 120v differential across its two wires as well as the nearly full amount of available X-10 signal, but when the light bulb is on, the signal is divided and the dimmer unit will not receive quite as much. Although this situation is rare, it happens occasionally when the signal level is marginal to begin with.
This effect is greatly exaggerated when a two-wire dimmer is used with a
non-linear load, like a low-voltage transformer (see figure 10) or a ceiling fan motor. I
know of many people who have attempted to use a two-wire dimmer with a transformer and
discovered that if it works at all, it is very intermittent. Either it will not stay on,
will not stay off, or they can not turn it on and/or off remotely.
With all of these potential problems what then are you supposed to
do? Well, there are ways to get around most of these problems. First, giving the dimmer
unit its own connection to line and neutral makes it possible to have a power source that
is unaffected by the on/off condition of the load (see figure 11). This does not mean that
somehow you can use a 2-wire dimmer by some wiring trick. This means that you will have to
purchase a real 3-wire dimmer. Using a 3-wire dimmer will make sure that the X-10 signals
do not have to squeeze through the load (the transformer, in this case) in order to get to
the dimmer unit. These 3-wire dimmers are usually constructed to operate with non-linear
loads. They have better circuitry that helps them withstand the severe current
fluctuations associated with inductive loads.
Should you wish to control a ceiling fan, the manufacturer of the fan should be contacted to be sure that their motor is compatible with "triac based" dimmers (don't ask them about X-10 dimmers, they usually don't know what you are talking about.). The one-speed fans do the best. If the fan you are trying to control is already a multi-speed fan, then set it to the highest speed using its pull-chain switch. Now the 3-wire dimmer can be used to control the speed through a wider range. (See figure 12.) Don't expect to get reliable low speed operation. There is so little torque at low speeds the fan will most likely stall. Also expect the fan to hum a little at the lower to middle speed where the sine-wave chopping is most severe. If you are lucky, the hum will be so slight it won't be noticeable. Surprisingly, it's the cheapest ceiling fans that usually do the best.
Home Automation Systems can supply you with 3-wire dimmers from Advanced Control Technologies, under the part number RD160. However, just because the RD160 is made for non-linear loads does not automatically mean that your non-linear load wants to be controlled by a triac based dimmer. We all know that we will damage a television or radio if we try to control them with a dimmer unit. Trying to control a television's picture brightness or a radio's volume with a dimmer is definitely not a good idea. The same is true for some "non-linear" loads like some fluorescent lights, low-voltage lighting systems and motors.
If X-10 dimmers cause inductive loads like motors and transformers to buzz and hum, why don't all speed controls? That is a complex question and I can only give you part of the answer. Some triac based dimmers are modified for use with inductive loads by adding capacitive loads (big capacitors) in parallel with inductive load. This makes the overall load appear more "linear" to the dimmer unit. These capacitors help absorb the wild voltage swings present at that moment in time where the triac "switches on" (see figure 13). This helps smooth out the power and so the motor (or transformer) is not so shocked by the abrupt transitions. There is a project in the works to develop a modified X-10 dimmer that will also have a capacitor dampening circuit.
In those cases where dimming is just not a good idea, a switch that uses a relay instead of a triac is required. A relay, of course, is either on or off: it isn't a variable output and as such, will deliver a full 100% of the sine wave to the load. These receivers are known as "hard-contact" receivers and are available from ACT as our part numbers RS120 (for a single unit) or RS100 (for a master unit that can be used in 3-way installations with a slave switch). Since these are industrial/commercial grade devices, both are rated at 20 amps. They will also require their own connection to line and neutral, and so, they will not work in wall boxes where there are only two wires.
Here ends "Which One Should I Use -
Part II" and I hope this has shed some light on the use of dimmers with fans and
Bob has been very generous in letting me write about whatever I want but I think I will ask your help for the next piece. I received a lot of great email from the non-technical article on "Lucky Lindy Meets My Grandmother" ( http://www.hometoys.com/htinews/feb97/articles/kingery/lindy.htm ), so much that because of it, I have been asked to speak at two upcoming conventions. Now I need your feedback. Please help me choose one of the following subjects for the continuing saga of "Which One Should I Use - Part III".
Will it be:
1. - Basic Coupling (Passive Couplers and Repeaters)
2. - Three and Four-Way Switch Circuits (Why are There so Many Ways to Wire Them)
3. - Noise and Filtering (What is Noise Pollution, What Causes It, How to Stop It)
Cast your vote by emailing me at firstname.lastname@example.org
Phillip Kingery is the X-10 Technical Support representative for Advanced Control Technologies, Inc., Indianapolis, Indiana and is a well known instructor of X-10 technical classes routinely held around the country. Upcoming classes include: Philadelphia, Indianapolis, Dallas, Los Angeles and Atlanta. His email address is: email@example.com